Enhancing the Adaptability of the Deep-Sea BacteriumShewanella piezotolerans WP3 to High Pressure and LowTemperature by Experimental Evolution under H2O2 Stress

Zhe Xie,a Huahua Jian,a Zheng Jin,a Xiang Xiaoa,b,c

aState Key Laboratory of Microbial Metabolism, School of Life Sciences and Biotechnology, Shanghai Jiao TongUniversity, Shanghai, China

bState Key Laboratory of Ocean Engineering, School of Naval Architecture, Ocean and Civil Engineering,Shanghai Jiao Tong University, Shanghai, China

cLaboratory for Marine Biology and Biotechnology, Qingdao National Laboratory for Marine Science andTechnology, Qingdao, China

ABSTRACT Oxidative stresses commonly exist in natural environments, and microbeshave developed a variety of defensive systems to counteract such events. Althoughincreasing evidence has shown that high hydrostatic pressure (HHP) and low tem-perature (LT) induce antioxidant defense responses in cells, there is no direct evi-dence to prove the connection between antioxidant defense mechanisms and theadaptation of bacteria to HHP and LT. In this study, using the wild-type (WT) strainof a deep-sea bacterium, Shewanella piezotolerans WP3, as an ancestor, we obtaineda mutant, OE100, with an enhanced antioxidant defense capacity by experimentalevolution under H2O2 stress. Notably, OE100 exhibited better tolerance not only toH2O2 stress but also to HHP and LT (20 MPa and 4°C, respectively). Whole-genomesequencing identified a deletion mutation in the oxyR gene, which encodes the tran-scription factor that controls the oxidative stress response. Comparative transcrip-tome analysis showed that the genes associated with oxidative stress defense, an-aerobic respiration, DNA repair, and the synthesis of flagella and bacteriophage weredifferentially expressed in OE100 compared with the WT at 20 MPa and 4°C. Geneticanalysis of oxyR and ccpA2 indicated that the OxyR-regulated cytochrome c peroxi-dase CcpA2 significantly contributed to the adaptation of WP3 to HHP and LT. Takentogether, these results confirmed the inherent relationship between antioxidant de-fense mechanisms and the adaptation of a benthic microorganism to HHP and LT.

IMPORTANCE Oxidative stress exists in various niches, including the deep-sea ecosys-tem, which is an extreme environment with conditions of HHP and predominantly LT.Although previous studies have shown that HHP and LT induce antioxidant defense re-sponses in cells, direct evidence to prove the connection between antioxidant defensemechanisms and the adaptation of bacteria to HHP and LT is lacking. In this work, usingthe deep-sea bacterium Shewanella piezotolerans WP3 as a model, we proved that en-hancement of the adaptability of WP3 to HHP and LT can benefit from its antioxidantdefense mechanism, which provided useful insight into the ecological roles of antioxi-dant genes in a benthic microorganism and contributed to an improved understandingof microbial adaptation strategies in deep-sea environments.

KEYWORDS Shewanella, deep sea, experimental evolution, high pressure, lowtemperature, oxidative stress, stress adaptation

Oxidative stresses commonly exist in natural environments, and microbes havedeveloped a variety of defensive systems to counteract such events. HHP (high

hydrostatic pressure) and LT (low temperature) are two of the most remarkable

Received 1 November 2017 Accepted 10December 2017

Accepted manuscript posted online 21December 2017

Citation Xie Z, Jian H, Jin Z, Xiao X. 2018.Enhancing the adaptability of the deep-seabacterium Shewanella piezotolerans WP3 tohigh pressure and low temperature byexperimental evolution under H2O2 stress.Appl Environ Microbiol 84:e02342-17. https://doi.org/10.1128/AEM.02342-17.

Editor Robert M. Kelly, North Carolina StateUniversity

Copyright © 2018 American Society forMicrobiology. All Rights Reserved.

Address correspondence to Xiang Xiao,xoxiang@sjtu.edu.cn.

ENVIRONMENTAL MICROBIOLOGY

crossm

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 1Applied and Environmental Microbiology

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

characteristics of the majority of deep-sea environments (1). Both characteristics canaffect cellular function by negatively influencing cell integrity, membrane fluidity,macromolecular interactions, and energy utilization, which can lead to metabolicdisorders and redox imbalances (2, 3). As bacterial metabolism and cellular integrity aremaintained by balancing the redox state of all cellular components for optimal overallfunction, the disturbance of this balance is accompanied by an increase in oxidativestress (4). For example, as reactive oxygen species (ROS) might come from the electrontransport chain in the cell membrane, damage to the membrane caused by HHP couldlead to the leakage of electrons, which can be intercepted by O2 and form ROS. InShewanella benthica KT99 and Shewanella violacea DSS12, a novel membrane-boundccb-type quinol oxidase displaces the bc1 complex and cytochrome c oxidase to overcomeoxidative stress under HHP conditions. Moreover, LT could increase the solubility of oxygenand result in increased concentrations of ROS and, thus, the potential for oxidative damage(3). Therefore, bacteria living in such environments have to overcome oxidative stress.

Antioxidative genes/proteins are commonly found in marine extremophiles (5, 6). S.violacea DSS12, a psychrophilic facultative piezophile isolated from deep-sea sedi-ments, has approximately 15 types of antioxidative genes. A cold-adapted bacterium,Colwellia psychrerythraea 34H, has been isolated from Arctic marine sediments and hasapproximately 13 types of antioxidative genes. A series of recent studies revealed thatLT or HHP could lead to microbial oxidative stress responses and the expression ofantioxidant enzymes. The transcription of the oxidative stress-related genes ahpC, katG,oxyR, soxS, sodA, sodC, and grxA was found to be significantly induced in Salmonellaenterica serovar Typhimurium in response to cold stress (5°C) (7). Proteomic analysis ofthe responses of Lactococcus piscium CNCM I-4031 to cold stress showed that twoantioxidant proteins, Ahp and thioredoxin (Trx), were induced (8). Treatment of Esch-erichia coli O157:H7 with HHP (100 MPa) resulted in upregulated transcription levels ofthe antioxidative genes trxA, trxC, and grxA. dps, trxA, and trxB mutants of E. coli K-12were shown to be significantly more sensitive to HHP (400 MPa) than their wild-type(WT) counterparts (9). In addition, E. coli MG1655 mutants lacking ROS scavengers (KatE,KatF, SodAB, and SoxS) were also found to be more pressure sensitive than thewild-type strains (10). Moreover, a radioresistant bacterium, Deinococcus radioduransR1, which can resist oxidative stress, was also found to be highly piezo resistant (11). Allof those studies indicated that antioxidation might play a role in HHP and LT adapta-tion. In Saccharomyces cerevisiae, it has been proven that enhanced oxidative stresstolerance leads to enhanced tolerance to HHP stress (12). However, whether bacteriawith increasing antioxidant defense capacities would better adapt to double stresses ofHHP and LT remains unknown.

Shewanella strains are among the most abundant Proteobacteria in the deep-seaenvironment. Shewanella piezotolerans WP3 (here referred to as WP3) was isolated fromWest Pacific sediments at a depth of 1,914 m, where the in situ temperature is �2°Cand the hydrostatic pressure is 20 MPa (13). In this study, the associations ofantioxidants with adaptation to HHP and LT were investigated by developing anevolved strain, OE100, which was cultured with a challenge concentration of H2O2.Remarkably, OE100 exhibited increased H2O2 tolerance and simultaneously bettergrowth at 20 MPa and 4°C. Resequencing of the genome revealed that deletionmutations occurred in a key oxidative stress response transcriptional regulator,OxyR, and two polar flagellar genes. Transcriptional analysis revealed that the genesassociated with oxidative stress defense, anaerobic respiration, DNA repair, and thesynthesis of flagella and bacteriophage were differentially expressed in OE100compared with the WT at 20 MPa and 4°C. Through the construction of an OxyRin-frame deletion strain (ΔoxyRc), we confirmed that reduced OxyR levels couldcontribute to the adaptation of WP3 to HHP and LT. Taken together, the results ofour study suggest that oxidative stress defense is a common adaptation strategy forbacteria coping with cold deep-sea environments.

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 2

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

RESULTSObtaining the H2O2-tolerant WP3 strain by experimental evolution. Six bacterial

populations of the WP3 wild-type strain were evolved under H2O2 stress as describedin Materials and Methods. After 500 generations of growth, 6 single-colony isolates,named E1-1, E2-1, E3-1, E4-1, E5-1, and E6-1, from each evolved population wererandomly picked by plating and single-colony isolation. Among the six strains, E6-1 hadthe highest endpoint culture density (optical density at 600 nm [OD600] of 4.7; P � 0.05)in medium with 0.6 mM H2O2 (see Fig. S1 in the supplemental material). This strain wasnamed OE100 and selected for follow-up analysis.

The tolerance of evolved strain OE100 to H2O2 is significantly increased. Tocompare the levels of H2O2 tolerance of OE100 and the WT, viability assays, discdiffusion assays, and plating defect assays were performed. Although there was nosignificant difference in viability between OE100 and the WT with a low H2O2 concen-tration (1 mM), OE100 showed a higher survival rate than that of the WT when the H2O2

concentration was increased from 1 mM to 6 mM. When the H2O2 concentration was6 mM, the survival rate of OE100 was approximately 5 times higher than that of the WT(Fig. 1A). Moreover, the disc diffusion assay showed that the diameter of the inhibitionzone of OE100 was smaller than that of the WT, indicating that OE100 has a greaterability to resist H2O2 stress than the WT (Fig. 1B). In addition, the viability of the OE100strain was significantly improved, as demonstrated by plating defect assays (Fig. 1C). Allof these data collectively led to the conclusion that OE100 was more resistant to H2O2

than its ancestor WT strain, indicating that the antioxidation property of OE100 wassignificantly increased during the experimental evolution process.

FIG 1 Characteristics of OE100 in response to H2O2 stress. (A) Survival assay. H2O2 was added to mid-log-phase cultures to the finalconcentrations indicated. After 20 min, the samples were properly diluted and plated onto 2216E plates. Colony counting wasperformed after 48 h. (B) Disc diffusion assay. Paper discs loaded with 10 �l H2O2 at different concentrations were placed onto bacteriallawns (grown for 8 h). The results shown are from 48 h after the discs were in place. For data presentation, the sensitivity of each strainwas calculated by its average diameter (n � 4). Error bars represent standard deviations. (C) Plating defect assay. Mid-log-phasecultures were adjusted to 108 CFU/ml and diluted by 10-fold series dilutions, and 5 �l of each dilution was plated onto 2216E plateswith H2O2 added at different concentrations. Images were captured 48 h after plating. All experiments were performed three times,and similar results were obtained.

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 3

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

OE100 shows better growth at 4°C and 20 MPa. To test whether the increase inthe antioxidant defense capacity can enhance the adaptability of the WT to HHP andLT, the growth of OE100 at different pressures and temperatures was examined. Underambient pressure conditions (0.1 MPa), OE100 exhibited a higher growth yield than thatof the WT at both 20°C and 4°C (Fig. 2A). A similar result was obtained when OE100 wascultured at high pressure. The growth rates of OE100 and the WT were 0.08 h�1 and0.06 h�1, respectively, and OE100 showed a significantly higher cell density than thatof the WT at 4°C and 20 MPa (Fig. 2B), which are the in situ temperature and hydrostaticpressure in locations inhabited by WP3.

Genomic changes and natures of mutated genes. To identify the genetic basis ofthe increased fitness of the evolved strain under HHP and LT conditions, the genomesof H2O2-evolved strain OE100 and the ancestral WT strain were sequenced, and thedata were analyzed. We obtained 90,393 subreads with 173� coverage of the WTgenome and 181,030 subreads with 213� coverage of the OE100 genome. Thegenomic sequence of OE100 differed from the ancestral WT sequence by 5 deletions(Table 1) and 7 single-nucleotide polymorphisms (SNPs) (Table 2).

Interestingly, only one of the sequenced mutations occurred in a gene with awell-known role in the oxidative stress response. A 69-bp deletion occurred in theoxidative stress transcriptional regulator gene oxyR. OxyR features a sensory cysteineresidue that is directly oxidized by H2O2, thereby generating a disulfide bond that altersits binding to DNA and controls the transcription of H2O2 scavenger genes (14). Weconfirmed that the reduced amino acid sequence of OxyR in OE100 belonged to theC-terminal substrate-binding domain containing these two conserved cysteine residues(Cys203 and Cys212 in WP3 OxyR are homologous to the Cys199 and Cys208 in E. coliOxyR) (see Fig. S2 in the supplemental material).

In addition, a 1,899-bp deletion was located within the coding region of two fliC1genes (encoding polar flagellin). The swimming motility assay with OE100 showed thatthe deletion in fliC1 led to a significant decrease in the swimming ability (Fig. S3).Moreover, a single-base deletion and a 12-bp deletion in two separate genes of

FIG 2 Growth curve of OE100 in 2216E medium at different temperatures and pressures. (A) Growth of OE100 at0.1 MPa and different temperatures. (B) Growth of OE100 at 20 MPa and different temperatures. The average valuesand standard deviations, displayed by the error bars, resulted from three replicates. All of the data shown representresults from at least three independent experiments.

TABLE 1 Deletions present in OE100 compared to the WT

GenePosition ongenome Annotated function

Length ofdeletion (bp)

SWP_RS16075 3815802 Hypothetical protein 1SWP_RS09080 2105025 Hypothetical protein 12oxyR 4012375 Transcriptional regulator, LysR family OxyR 69fliC1a 1555674 FliC, polar flagellin domain protein 70fliC1b 1556355 FliC, polar flagellin domain protein 1,215

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 4

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

unknown function (Table 1) indicated the previously unrecognized mechanisms of theoxidative stress response and adaptation under experimental conditions.

SNPs were located within the coding regions of six genes (trkH, 23S rRNA, yjjW, acrR,rpoD, and tauE) with annotated functions and in one gene of unknown function(SWP_RS13360) (Table 2); no mutations were found in the noncoding or intergenicregions. Among these SNPs, there are only two missense mutations (yjjW and rpoD) andone nonsense mutation (SWP_RS13360). The yjjW gene encodes a glycine radicalenzyme activase that acts on a wide variety of biomolecules in numerous pathways,including vitamin biosynthesis and coenzyme biosynthesis (15). The rpoD gene encodesa �70 protein that can affect the expression state of cells (16). Therefore, mutations inthese two genes could rewire the network into a more efficient state and affect thefitness of OE100.

Transcriptomic changes in OE100 compared to the WT. To check the contribu-tion of the other transcriptional rearrangements to the increased fitness of OE100 atHHP and LT, differentially expressed genes between the WT and OE100 at maximumgrowth yields at 20 MPa and 4°C were investigated by whole-genome transcriptionalanalysis. To validate the transcriptome sequencing (RNA-Seq) data, 9 genes, includingthose that were upregulated, downregulated, or unchanged, were selected for reversetranscription (RT)-quantitative PCR (qPCR). The relative mRNA levels for each gene werecalculated and log transformed. The correlation coefficient (R2) between the dataobtained by microarray analysis and those obtained by qPCR was 0.9213, demonstrat-ing that the transcriptome data were reliable and could be used for the follow-upanalysis (see Fig. S4 in the supplemental material).

Compared to the WT, 483 genes (approximately 10% of the WP3 genome) in OE100showed significant differential expression (log2 fold change for OE100 compared to theWT of �1; P value [false discovery rate {FDR}] of �0.001), of which 302 genes and 181genes were up- and downregulated, respectively (Fig. S5). Specifically, the transcriptionlevels of most genes encoding the scavenger enzymes involved in hydrogen peroxideor peroxide elimination were upregulated (Table 3). For example, katE and ahpC, whichare genes that encode the primary H2O2 scavengers in many bacteria, showed 4-fold-and 32-fold-higher expression levels in OE100 than in the WT, respectively. ccpA2 andgpx, which are also responsible for H2O2 scavenging, were upregulated by 1.6-fold and3.6-fold, respectively. In addition, the transcription level of dps, which is a homologueof the dps gene in E. coli with both the “ferritin-like” activity and the DNA-bindingcapacity, was upregulated by 1.68-fold. However, the transcription levels of the genesinvolved in superoxide scavenging (sod) were not significantly different.

In addition, most of the genes related to anaerobic respiration on different electronacceptors, including nitrate, nitrite, fumarate, and dimethyl sulfoxide (DMSO), showedsignificant upregulation in OE100, while the genes related to the aerobic electrontransport chain were downregulated. CymA, a cytoplasmic-membrane-bound c-typecytochrome belonging to the NapC/NirT protein family that directs electrons from themenaquinol pool to the periplasm, was induced by 1.7-fold in OE100. This protein wasproven to be the essential electron transport protein for the anaerobic respiratoryflexibility of the Shewanellae (17). In WP3, CymA is involved in the reduction of nitrate,nitrite, DMSO, and fumarate, and it was proven to serve as the sole electron transporter

TABLE 2 SNPs present in OE100 compared to the WT

Gene Position on genome Annotated functionType of amino acid change(amino acid change) Nucleotide change

trkH 56774 K� transporter Trk Silent (S483S) G¡A23S rRNA 76729 23S rRNA T¡CyjjW 228972 Pyruvate formate lyase-activating enzyme Missense (H42Q) C¡GacrR 244302 TetR family transcriptional regulator Silent (L84L) C¡TrpoD 1115319 RpoD RNA polymerase sigma factor 70 Missense (A273G) C¡GSWP_RS13360 3125851 Hypothetical protein Nonsense (Q464Stop) T¡CtauE 4099249 Cytochrome c biogenesis Silent (A326A) A¡C

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 5

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

TABLE 3 Selected differentially expressed genes in OE100 at 4°C and 20 MPa (probability � 0.8)

Function Gene Gene descriptiona

Log2 fold change(OE100/WT)

Antioxidant activity SWP_RS05640 Glutathione peroxidase Gpx 1.85SWP_RS16945 Transcriptional regulator OxyR 1.55SWP_RS00790 DNA-binding protein Dps 0.75SWP_RS19255 Transcriptional regulator OhrR 1.80SWP_RS19285 Alkyl hydroperoxide reductase subunit F, AhpF 1.24SWP_RS19290 Alkyl hydroperoxide reductase subunit C, AhpC4 1.62SWP_RS19745 Catalase KatE 2.05SWP_RS12910 Cytochrome c peroxidase CcpA2 0.69

Nitrate respiration SWP_RS19850 Anaerobic dehydrogenases NapA� 1.86SWP_RS19855 Nitrate reductase NapB� 1.95SWP_RS19845 Periplasmic nitrate reductase NapD� 1.69

Nitrite respiration SWP_RS20710 Formate-dependent nitrite reductase, NrfC protein 6.50SWP_RS20715 Formate-dependent nitrite reductase, NrfG protein 7.07SWP_RS20705 Formate-dependent nitrite reductase, NrfD protein 5.87SWP_RS17675 Nitrite reductase 1.92

Anaerobic respiration SWP_RS21330 Membrane-bound tetraheme cytochrome c CymA 0.78

Fumarate respiration SWP_RS01965 Fumarate reductase cytochrome b subunit 1.69SWP_RS01970 Fumarate reductase respiratory complex subunit 1.82SWP_RS01975 Fumarate reductase flavoprotein subunit 1.60SWP_RS01980 Succinate dehydrogenase/fumarate reductase iron-sulfur protein 1.58

DMSO respiration SWP_RS03195 Decaheme cytochrome c CytC 1.98SWP_RS03200 Transcriptional regulator, LysR family, LysR2 1.65SWP_RS03205 Twin-arginine translocation pathway signal, DmsA2 1.98SWP_RS03215 Cytoplasmic chaperone TorD 1.73SWP_RS03210 4Fe-4S ferredoxin, iron-sulfur binding 1.82SWP_RS03220 Hypothetical protein 1.37SWP_RS15435 Conserved hypothetical protein DmsH 1.16

Polar flagellum SWP_RS06720 Flagellin FlaG1 1.89SWP_RS06655 Flagellar basal body rod protein FlgB1 1.33SWP_RS06660 Flagellar basal body rod protein FlgC1 1.23SWP_RS06775 ATPase FliI/YscN �1.02SWP_RS06690 Flagellar P-ring protein �1.03SWP_RS06770 Flagellar assembly protein FliH1 �1.17SWP_RS06780 Flagellar protein FliJ1 �1.34SWP_RS06715 Flagellin, FliC1b �3.07

Lateral flagella SWP_RS22555 Sodium-type flagellar protein MotY 1.83SWP_RS22550 Motor switch part FliM 1.39SWP_RS22700 Flagellin, FliC2 �3.82

Phage SWP_RS11565 Putative capsid protein FpsC �5.58SWP_RS11550 Minor capsid protein FpsD �9.92SWP_RS11570 Zonula occludens toxin FpsH �5.47SWP_RS11545 ssDNA-binding protein FpsB �9.32SWP_RS11560 Minor capsid protein FpsF �6.31SWP_RS11535 Transcription regulator FpsR �6.31SWP_RS11540 Replication protein FpsA �11.22

DNA repair SWP_RS05950 DNA repair protein RecN �2.35SWP_RS23035 DNA polymerase ImuA �3.04SWP_RS10395 DNA polymerase ImuB �2.79SWP_RS10400 DNA polymerase V DnaE2 �2.47SWP_RS05205 DNA polymerase IV �3.02SWP_RS21400 Peptidase S24, LexA repressor �2.04SWP_RS06045 DNA recombination/repair protein RecA �0.87SWP_RS20905 MutT/nudix family protein �1.99

assDNA, single-stranded DNA.

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 6

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

toward nitrite reductase (18). Moreover, compared to other anaerobic respirationpathways, nitrite reductases have the highest expression levels (nrfC, 89-fold; nrfG,134-fold; nrfD, 58-fold).

Most of the genes involved in flagellar biosynthesis, phage production, and DNArepair were repressed (Table 3). The genes encoding polar flagellin (fliC1b) and lateralflagellin (fliC2) were downregulated by 8.4-fold and 14-fold, respectively. The genesencoding filamentous phage capsid proteins (FpsC, FpsD, and FpsF) were significantlydownregulated by 47.8-fold, 968.7-fold, and 79.3-fold, respectively. The expression levelof DNA polymerase IV, which is a translesion polymerase known to promote damagedDNA survival by allowing replication to bypass lesions, was decreased by 8.1-fold. Inaddition, the transcription levels of two key genes involved in the SOS response, lexAand recA, were repressed by 4.1-fold and 1.8-fold, respectively.

Deletion mutation of oxyR results in increased growth of WP3 at 20 MPa and4°C. There was a 69-bp deletion located in the C-terminal substrate-binding domain ofoxyR in OE100. To confirm whether the reduced level of the oxyR gene was involved inLT and HHP adaptation, an OxyR in-frame deletion strain (ΔoxyRc) with an identical69-bp deletion was constructed based on WP3Δ4RE, which was a wild-type WP3 strainwith 4 restriction enzyme gene deletions and a higher conjugation efficiency (19). TheΔoxyRc strain carrying pSW2 showed a higher tolerance to H2O2 than did WP3Δ4RE/pSW2, and this phenotype was abolished in the ΔoxyRc/oxyR strain (complementationof the ΔoxyRc strain with the oxyR gene) (Fig. 3A), suggesting that the mutation in OxyRwas responsible for the phenotype. Notably, the ΔoxyRc strain carrying pSW2 demon-strated better growth at 20 MPa and 4°C, indicating that reduced oxyR levels in OE100contributed to the growth of OE100 at 20 MPa and 4°C (Fig. 3B). Interestingly, out of 30single-colony isolates that were randomly picked from 6 evolved populations (5single-colony isolates from each evolved population), 19 strains were found to harborthe same deleted region in oxyR as the one present in OE100 (see Table S2 in thesupplemental material), indicating that the mutation of this region in oxyR is importantfor WP3 survival under H2O2 stress.

As OxyR regulates different kinds of antioxidant genes, to identify the OxyR-regulated genes involved in adaptation to HHP and LT, the relative mRNA levels ofthese genes in the ΔoxyRc strain carrying pSW2 at 20 MPa and 4°C were examined byRT-qPCR (Fig. 3C). The transcription of the oxyR gene was upregulated in the ΔoxyRcstrain carrying pSW2, suggesting that self-repression is present in oxyR expression.Expectedly, the transcriptions of the katE, ahpF, and ahpC4 genes were significantlyupregulated in the ΔoxyRc strain carrying pSW2 and downregulated in the ΔoxyRc/oxyRstrain, indicating that OxyR in WP3 acts as a repressor of the katE, ahpF, and ahpC4genes, and the absence of a conserved domain stimulates the expression of thesegenes. Notably, the transcription of the ccpA2 gene was upregulated in both the ΔoxyRcstrain carrying pSW2 and the ΔoxyRc/oxyR mutant, thus demonstrating a similar patternof change with oxyR and suggesting that OxyR acts as an activator of the ccpA2 gene.Accordingly, the ΔoxyRc/oxyR mutant with a significantly high level of expression of theccpA2 gene also demonstrated better growth at 20 MPa and 4°C (Fig. 3B), indicatingthat the ccpA2 gene was important for adaptation to HHP and LT. To test this notion,we constructed ccpA2 deletion (ΔccpA2/pSW2), complementation (ΔccpA2/ccpA2), andoverexpression (WT strain carrying ccpA2) mutants. Compared to the WT, the ΔccpA2strain carrying pSW2 showed decreased growth at HHP and LT, while growth wasincreased in the ΔccpA2/ccpA2 strain and in the WT strain carrying ccpA2 (Fig. 3D).Moreover, the ΔccpA2 strain carrying pSW2 did not show a growth defect at 0.1 MPaand 20°C (Fig. S6), suggesting that ccpA2 is the major antioxidant facilitating theadaptation of WP3 to HHP and LT.

DISCUSSION

Based on our results, we have proposed a model to explain the enhanced HHP/LTadaptation mechanism of OE100 (Fig. 4). The most important system that contributesto adaptation to HHP and LT is an antioxidant defense system. As the substrate-binding

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 7

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

domain of oxyR in OE100 was mutated, we inferred that OxyR lost the ability to senseH2O2 and affect its binding to target sites. OE100 showed greater resistance to H2O2,indicating that OxyR in OE100 functioned as a repressor for H2O2 scavenger genes, andthe mutation of oxyR derepressed the expression of these genes. Through the con-struction of the ΔoxyRc mutant, we found that the ΔoxyRc mutant demonstrated betteradaptation to HHP and LT with an increased antioxidant ability, indicating that anti-oxidation indeed contributes to adaptation to HHP and LT. The upregulated genes katE,ahpF, ahpC4, and ccpA2 in the ΔoxyRc strain were also found to be induced in OE100at 20 MPa and 4°C, suggesting that these genes were under the control of OxyR andcontribute to adaptation to HHP and LT. Complementation with OxyR did not restore

FIG 3 Characteristics of the ΔoxyRc strain. (A) Disc diffusion assay of the ΔoxyRc strain carrying pSW2 and the ΔoxyRc/oxyR strain inresponse to H2O2. (B) Growth curves of the ΔoxyRc strain carrying pSW2 and the ΔoxyRc/oxyR strain at 20 MPa and 4°C. (C) Impact ofreduced OxyR levels on expression levels of various H2O2-scavenging genes. The expression levels of these genes in the ΔoxyRc straincarrying pSW2 and the ΔoxyRc/oxyR strain were normalized relative to those in WP3Δ4RE/pSW2. (D) Impact of the ccpA2 gene on thegrowth of WP3 at 20 MPa and 4°C. All experiments were performed three times, and similar results were obtained.

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 8

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

the wild-type phenotype at 20 MPa and 4°C. Previously, the copy numbers of pSW2 inWP3 were determined (12 and 30 copies at 20°C and 4°C, respectively) (20). Thetranscription level of oxyR was upregulated by 18-fold in the complemented straincompared with the wild-type strain (Fig. 3C). Accordingly, the mRNA level of ccpA2 wassignificantly higher in the oxyR-complemented strain than that in the wild-type strain.Although the transcription levels of other antioxidant genes (katE, ahpC1, ahpC2,ahpC3, ahpC4, and ahpF) were significantly downregulated (Fig. 3C), the increasedccpA2 expression level still conferred a higher antioxidant ability to the complementedstrain than that of the wild-type strain (Fig. 3B). Therefore, of these genes, ccpA2 wasindicated to be the most important gene for adaptation to HHP and LT. Notably, CcpA2is a homologue of CcpA in Shewanella oneidensis MR-1, which was positively regulatedby the OxyR protein in Shewanella oneidensis MR-1 (OxyRso) and functions as a majorH2O2-decomposing enzyme under anoxic conditions (21, 22). Recent findings showedthat CcpA allowed bacteria to use H2O2 as a terminal electron acceptor to supportanoxic respiration (23). AhpF, which is a flavoprotein with NADH:disulfide oxidoreduc-tase and functions as an assistant to restore the disulfide in AhpC4 to its reduced form(24), was highly induced in the ΔoxyRc strain, indicating a demand for reducingequivalents of the strain for adaptation to HHP and LT. In addition to scavengerenzymes, the genes of the iron storage protein Dps were also upregulated in OE100. AsDps can sequester free iron inside the cell to prevent the excessive generation ofharmful HO· by Fenton’s reaction and shield DNA from damage (25), the upregulationof dps in OE100 suggested a need to minimize free iron during adaptation to HHPand LT.

In addition to an antioxidant defense system, an energy production system alsocontributes to adaptation to HHP and LT. WP3 has two sets of nitrate respirationsystems, periplasmic dissimilatory nitrate reductase alpha (NAP-�) and NAP-�. A pre-vious study demonstrated that both systems were functional, but NAP-� was a moreefficient nitrate-reducing system (18). In addition, two sets of DMSO respiration systemswere also identified in WP3, and both DMSO type I and type II systems were shown tobe functional, but type I was more efficient for WP3 to thrive at 20 MPa and 4°C (26).In OE100, the NAP-� and DMSO type II systems showed higher expression levels inOE100 than in the WT, indicating that OE100 strengthens the less efficient respirationsystems and thus produces more energy.

FIG 4 Graphic display of transcriptomic data for genes categorized by function in OE100 at 20 MPa and 4°C. Green andred indicate the genes that are induced and repressed (OE100/WT), respectively. Underlining indicates that the gene ismutated in OE100. Blue arrows represent gene regulation. The red rectangle indicates that the gene is important foradaptation to HHP and LT. DMS, dimethyl sulfide.

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 9

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

Moreover, the better adaptation of OE100 to HHP and LT can be partly due torepressed systems, including the SOS response and DNA repair systems. Althoughneither H2O2 nor O2

� can damage DNA directly, H2O2 produces HO·, which can oxidizeboth base and ribose moieties of the DNA, giving rise to a wide variety of lesions andthus inducing the SOS response and activating the expression of DNA repair genes (24).Stress survival studies showed that an SOS-deficient mutant of Listeria monocytogeneswas less resistant to H2O2 than the wild type (27). Compared to the WT, the key genesinvolved in the SOS response, lexA and recA, were downregulated in OE100, indicatingthat DNA damage under the tested conditions was not serious enough to activate theSOS response. In addition to recA and lexA, the transcription levels of some genesencoding DNA repair enzymes, including ImuA/B, MutT, and DNA polymerase IV, alsoshowed downregulation, suggesting a low level of DNA damage that may be due to theenhanced antioxidation properties of OE100. Therefore, OE100 can redirect less energyutilization from proliferative processes to macromolecular processes and repair.

WP3 had two sets of flagellar systems, and a previous study showed that theexpression of polar and lateral flagellar systems could respond to HHP and LT, respec-tively. The mutant lacking lateral flagella showed a significant decrease in growth at4°C, which indicated that lateral flagella are important for the environmental adapta-tion of the strain at LT (28). Flagellins constitute the major extracellular part of the polarflagellum, which is responsible for swimming motility. In OE100, the major extracellularpart of the polar flagellum of OE100 was lost during experimental evolution, leading toa decreased swimming motility of OE100 (see Fig. S3 in the supplemental material).Interestingly, out of 30 single-colony isolates that were randomly picked from 6 evolvedpopulations, 6 strains were found to harbor the same deleted region in the polarflagellar gene fliC1, and all 6 of these isolates showed decreased swimming motility(Table S2). A trade-off between motility and oxidative stress resistance was also foundfor Pseudomonas putida KT2440, where the nonflagellated strain showed better resis-tance to paraquat-induced oxidative stress (29). Therefore, OE100 with impaired polarflagella can better cope with oxidative stress induced by HHP and LT. Moreover, aprevious study demonstrated that the filamentous phage SW1 was active at 20 MPaand 4°C, and a significantly higher growth rate was observed for the phage-free strainthan for wild-type strain WP3 at 20 MPa and 4°C (19, 30). Therefore, the downregulationof phage genes may partly contribute to the better growth ability of OE100. As thesynthesis of both flagella and phage require plenty of energy, the repression of thesesystems showed that OE100 had more energy available to better adapt to HHP and LT.

Furthermore, a previous study showed that the overexpression of a single antioxi-dant enzyme of Kat, Ahp, or Dps could help bacteria to cope with many environmentalstresses (e.g., cold, acid, heat, heavy metal, high salt concentrations, and UV) which canlead to oxidative stress (31–38), suggesting that antioxidant defense is a commonadaptation strategy used by bacteria in response to environmental stressors other thanHHP and LT.

In summary, this study confirmed the inherent relationship between antioxidantdefense mechanisms and adaptation to HHP and LT, provides new scientific guidancefor adaptation mechanisms of deep-sea bacteria, and provides an alternative methodto obtain a bacterium with enhanced tolerance to HHP and LT by increasing itsantioxidation capacity through experimental evolution. In the future, the protection ofcells from oxidative stress damage should be taken into account when mimicking thecultivation of deep-sea bacteria or studying the diversity or characteristics of uncultur-able deep-sea microorganisms. Furthermore, this study provides insight into the pres-sure adaptation mechanism of Shewanella, and the transcriptomic data obtained fromthis study provide a valuable resource for further functional studies of the underlyingmolecular mechanisms in deep-sea bacteria.

MATERIALS AND METHODSExperimental evolution. All bacterial strains and plasmids used in this study are listed in Table

4. The experimental evolution of WP3 under external H2O2 stress was performed as previously

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 10

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

described, with some modifications (39–41). Briefly, a single clonal isolate of Shewanella piezotoler-ans WP3 was used as the ancestor (WT) to 6 populations for experimental evolution. Six populationswere propagated in 50 ml 2216E medium (5 g/liter tryptone, 1 g/liter yeast extract, and 34 g/literNaCl). The cell cultures were aerobically grown at 20°C and at 200 rpm and were serially transferredto 50 ml 2216E medium with 0.4 mM H2O2 every 12 h with 1:100 dilutions. After 250 generations,the concentration of H2O2 added to the dilution medium was adjusted to 0.6 mM, and thepopulations evolved for 250 more generations. After a cumulative 500 generations of evolution, 6single-colony isolates from each evolved population were randomly picked by plating and single-colony isolation. As one measure of evolving fitness, the endpoint growth level (culture density atthe stationary phase) was measured in the cultures derived from each isolate. The strain with thehighest endpoint culture density was selected and designated OE100.

Growth assay. The growth of the ancestral strain (WT) and the evolved strain (OE100) wasexamined with three replicates for each strain. For aerobic growth, the strains were cultured in2216E broth in a rotary shaker (200 rpm) at 20°C or 4°C. For the anaerobic growth assay, serumbottles, each containing 100 ml of fresh medium, were anaerobically prepared by flushing withnitrogen gas through a butyl rubber stopper fixed with a metal seal to strip the dissolved oxygenprior to autoclave sterilization. To examine the growth of the strains at HHP, each culture was grownto log phase in 2216E medium at 1 atm and 20°C or 4°C in a rotary shaker. The log-phase cultureswere diluted into the same medium to an optical density at 600 nm of 0.01. Aliquots of the dilutedculture were injected into serum bottles containing 100 ml of anaerobic medium through the butylrubber stopper. After brief shaking, 2.5-ml disposable plastic syringes were used to distribute theculture in 2-ml aliquots. The syringes with needles were stuck into rubber stoppers in a vinylanaerobic airlock chamber (Coy Laboratory Products Inc., Grass Lake, MI, USA). The syringes werethen incubated at a hydrostatic pressure of 20 MPa at 4°C or 20°C in stainless steel vessels (FeiyuScience and Technology Exploitation Co. Ltd., Nantong, China) that could be pressurized by usingwater and a hydraulic pump (5, 18, 26). These systems were equipped with quick-connect fittings forrapid decompression and recompression (achieving 20 MPa within 30 s). The cultures were held at20 MPa and 20°C and at 20 MPa and 4°C for 24 h and 84 h, respectively.

H2O2 challenge assay. The response of the bacterial strains to H2O2 challenge was assessed bymultiple approaches. A disc diffusion assay was employed to determine the impacts of H2O2 onbacterial lawns. Mid-log-phase cells were inoculated by 10-fold dilution into fresh broth and spreadonto fresh 2216E plates (200 �l of culture; approximately 4 � 106 CFU). After 8 h, 6-mm-diameterpaper discs loaded with 10 �l H2O2 (Sangon Biotech Co. Ltd., China) at concentrations of 1, 2, 4, and8 M were placed onto the bacterial lawn. Plates were incubated at 20°C for 2 days. To determine theeffect of H2O2 on the survival of cells, H2O2 at concentrations of 1, 2, 4, and 6 mM was added to themid-log-phase cultures. After 20 min of incubation at 20°C, treatment was stopped by addingcatalase (Worthington Biochemical Corporation, USA) at a concentration of 400 U/ml. The sampleswere diluted in a 3.4% NaCl solution and plated onto 2216E plates for counting viable colonies.

TABLE 4 Bacterial strains and plasmids used in this study

Strain or plasmid Description Reference or source

StrainsE. coli WM3064 Donor strain for conjugation; ΔdapA 46

S. piezotolerans WP3WT Wild-type strain Laboratory stockOE100 Evolved strain derived from WP3 This studyWP3Δ4RE Wild-type WP3 strain with 4 restriction enzyme gene deletions 19WT/pSW2 WT with pSW2 This studyWP3Δ4RE/pSW2 WP3Δ4RE with pSW2 This studyΔoxyRc oxyR mutant derived from WP3Δ4RE This studyΔoxyRc/pSW2 ΔoxyRc strain with pSW2 This studyΔoxyRc/oxyR ΔoxyRc strain with pSW2-oxyR This studyΔccpA2 ccpA2 deletion mutant derived from WT This studyΔccpA2/pSW2 ΔccpA2 with pSW2 This studyΔccpA2/ccpA2 ΔccpA2 with pSW2-ccpA2 This studyWT carrying

ccpA2WT with pSW2-ccpA2 This study

PlasmidspRE112 Chloramphenicol resistance; suicide plasmid with sacB1 gene as a negative selection

marker; used for gene deletion47

pRE112-ΔoxyRc pRE112 containing the PCR fragment for deleting oxyR This studypRE112-ΔccpA2 pRE112 containing the PCR fragment for deleting ccpA2 This studypSW2 Chloramphenicol resistance; generated from filamentous bacteriophage SW1; used for

complementation20

pSW2-oxyR pSW2 containing oxyR for complementation This studypSW2-ccpA2 pSW2 containing ccpA2 for complementation and overexpression This study

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 11

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

Colonies were counted after 2 days of incubation at 20°C. The results were expressed as percentsurvival, which is calculated by dividing the CFU per milliliter of surviving cells after 20 min by theCFU per milliliter of cells after 0 min. To examine the plating defects, cells at mid-log phase wereadjusted to approximately 108 CFU/ml with fresh 2216E medium, followed by six 10-fold serialdilutions. Five microliters of each sample was spotted onto 2216E plates, and H2O2 was added atconcentrations of 0.02, 0.04, and 0.08 mM. The plates were incubated for 2 days before being read.All experiments were repeated at least three times.

Swimming motility assay. The swimming motility assays were performed as previously described(42). Briefly, 1 �l of culture (mid-exponential phase) from each strain was placed onto swimming plates(2216E medium with 0.3% [wt/vol] agar). The plates were incubated at 20°C for 3 days. Motility wasassessed by measuring the diameter of the colony. The assay was performed at least three independenttimes and in triplicate.

Genome sequencing. Cells of OE100 and WT bacteria growing in the exponential phase wereharvested by centrifugation (12,000 � g for 3 min at 20°C), and the cell pellets were used for DNAextraction. High-molecular-weight DNA was extracted by using the phenol-chloroform method andsequenced by using the PacBio RS II sequencer with the P6-C4 sequencing reagent (Wuhan Institute ofBiotechnology, China). We used the sequenced and annotated GenBank file of the genome of S.piezotolerans WP3 (GenBank accession no. NC_011566) as a reference and mapped the sequencedgenomes.

Mutations in the evolved strain versus the ancestor were confirmed by the identification of SNPs anddeletion-insertion polymorphisms (DIPs) by using Mauve2.4.0 (Darling Laboratory, University of Tech-nology—Sydney). The presence of DIPs and suspected SNP sites in OE100 was verified by Sangersequencing. The primers designed to amplify the PCR products for Sanger sequencing are listed in TableS1 in the supplemental material. Differences found in our ancestral strain from the reference WP3genome were considered the “genetic property” of the ancestor.

Deletion mutagenesis and complementation. In-frame deletion mutagenesis of oxyR was per-formed as previously reported (26). The primers designed to amplify the PCR products for mutagenesisare summarized in Table 5. The PCR products were generated by using the DNA of OE100 as thetemplates with primers oxyRc-UF and oxyRc-UR. After digestion with the restriction enzymes SacI andKpnI, the treated fusion PCR products were ligated into the SacI and KpnI sites of suicide plasmid pRE112,resulting in mutagenesis vector pRE112-oxyRc. This vector was first introduced into E. coli WM3064 andthen conjugated into WP3Δ4RE. Positive exconjugants were spread onto marine agar 2216E platessupplemented with 10% (wt/vol) sucrose. The chloramphenicol-sensitive and sucrose-resistant colonieswere screened for the oxyRc deletion by PCR. The ΔoxyRc mutant was verified by sequencing of themutated region. The same strategy was used to construct the ΔccpA2 strain.

Plasmid pSW2 was used for the genetic complementation of the mutant. The intact oxyR genefragment was amplified from WT genomic DNA. The resulting PCR products were inserted into theshuttle vector pSW2. The ΔoxyRc/oxyR transformant was obtained by introducing recombinant plasmidpSW2-oxyR into the ΔoxyRc strain via conjugation. The same strategy was used to construct thecomplementing ΔccpA2/ccpA2 strain.

RNA sequencing and RNA-Seq data analysis. Total RNA from the WT and OE100 mutant strainsat maximum growth yield under 20 MPa and 4°C was extracted by using TRIzol reagent accordingto the manufacturer’s instructions (Invitrogen, Carlsbad, CA, USA). rRNA was removed by using aRiboMinus transcriptome isolation kit (Invitrogen, Carlsbad, CA, USA). Fragmentation buffer wasadded for interrupting mRNA into short fragments. Taking these short fragments as the templates,a random hexamer primer was used to synthesize the first-strand cDNA. The second-strand cDNAwas synthesized by using buffer, dATPs, dGTPs, dCTPs, dUTPs, RNase H, and DNA polymerase I afterthe removal of deoxynucleoside triphosphates (dNTPs). Short fragments were purified with theQIAquick PCR purification kit (Qiagen, Hilden, Germany) and resolved with elution buffer (EB; 10 mMTris-Cl, pH 8.5) for end repair and adding poly(A). After that, the short fragments were connectedwith sequencing adapters. The uracil-N-glycosylase (UNG) enzyme was then used to degrade thesecond-strand cDNA, and the product was purified by using a Mini Elute PCR purification kit (Qiagen,Hilden, Germany) before PCR amplification. RNA sequencing was performed on an Illumina HiSeq2000 instrument (Illumina, San Diego, CA, USA) at the Beijing Genomics Institute (BGI) (China). A totalof three independent biological replicates were prepared under each condition.

The clean data were trimmed from raw reads for further analysis by removing adapter sequences,low-quality reads (Q value of �20), ambiguous “N” nucleotides, and fragments of �20 bp. Uniquesequences were mapped to the WP3 genome with SOAPaligner (soap2) (43). The uniquely mappedreads were collected and analyzed with the DEG-Seq package to identify the differentially expressedgenes (estimation of gene expression levels based on the reads per kilobase per million [RPKM]values). Criteria of an FDR of �0.001 and the absolute value of a log2 fold change of �1 were usedas the thresholds to judge the significance of the gene expression differences. The NOISeqBIOmethod was used to handle the biological replicates, and a probability of �0.8 and a log2 foldchange of �1 were used as the thresholds to judge if the gene was differentially expressed underthe two experimental conditions (44).

RT-qPCR. To validate the data generated by RNA-Seq, the expression levels of 10 randomly selectedgenes were quantified by using RT-qPCR. The primers were designed by using Primer Express software(Applied Biosystems, CA, USA) and are listed in Table 5. The amount of target was normalized to thereference gene swp2079 (45). A RT-qPCR assay of each biological replicate was carried out for threereplicates, and the mean values and standard deviations were calculated for the relative RNA expression

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 12

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

levels. The RT-qPCR log2 ratio values were plotted against the RNA-Seq data log2 values, and thecorrelation coefficients (R2) between these two data sets were calculated.

In the assessment of the effect of reduced levels of OxyR on the expression levels of variousH2O2-scavenging genes, the expression levels of target genes in the ΔoxyRc strain carrying pSW2 and theΔoxyRc/oxyR strain were normalized relative to those in WP3Δ4RE/pSW2.

Accession number(s). The RNA-Seq raw reads were deposited in the Gene Expression Omnibus(GEO) database of the NCBI with accession no. GSE99297.

TABLE 5 Primers used in this study

Oligonucleotide Sequence (5=¡3=) Description

oxyRc-UF GGCCGGTACCTTATTTTGTATGGAACAAAATAGACCATTTAGTCG Gene deletionoxyRc-UR CTGTGAGCTCTGGTTAATTTACATCAAGAACAAAATTTTAATCGC Gene deletionoxyRF-C TCGACTCGAGATGTCATTGGTAATCTGCTTTATGAAAC ComplementationoxyRR-C CGCGGATCCTTAATTATGTTGTTCAAGCAGCTTTTG ComplementationccpA2UF TGCCGGTACCCGCTTGGTTGCTACTCCC Gene deletionccpA2UR ATAGTAAGGTTTGTCAATGTGCTC Gene deletionccpA2DF GACAAACCTTACTATGCCACGTCCTGTGCCTT Gene deletionccpA2DR GAACGAGCTCACAATACAGAGCCACACAACTGGTC Gene deletionccpA2F-C CCGCTCGAGATGAGCACATTGACAAACCTC ComplementationccpA2R-C CGCGGATCCCTAGTTAGCAAAAGGCACAG ComplementationoxyR-F GCAATTGGTATTAAAAACCCTAAATAC oxyR sequencingoxyR-R AAGGATGCCTTGAGAAATCTATG oxyR sequencingkatF CCTGCGTGACGGGATTAAGT RT-qPCRkatR CGGGTTGCGCTTTTGG RT-qPCRahpC1F CGCATGATCGACGCACTTC RT-qPCRahpC1R CCAACCAGCAGGACAAACATC RT-qPCRahpC2F CACTTCTAAGCGACGAAGACCAT RT-qPCRahpC2R AATTTCTTCTCCCCCCAAACA RT-qPCRahpC3F CCAAACTCACTGCCGGAAA RT-qPCRahpC3R GCACCAGCCACCACGATAG RT-qPCRahpC4F CCCAACTGGCGCAATCAC RT-qPCRahpC4R GCGCTAAGCCATCTTCTTCAA RT-qPCRahpFF AACGTACCGGGTGAGCAAGA RT-qPCRahpFR TCACAATGTGGGCAATAAGCA RT-qPCRccpA1F ACTCGAGAACGAATTGCACAAG RT-qPCRccpA1R CGGCGCGATAGGATGGT RT-qPCRccpA2F GGGTTGGCCCAGTTTTGG RT-qPCRccpA2R GACCGCCTGCTTGCTCTTTA RT-qPCRccpA3F GCCTCATGCCATGTGCAA RT-qPCRccpA3R ACCAATACTCAAGCGACGAGAGT RT-qPCRccpA4F TGCCCCAAGCCGCTATC RT-qPCRccpA4R CAATAAATGCAGCCCAAATCC RT-qPCRgpxF GGTGCTCAAGAAAAAGGCGATA RT-qPCRgpxR CCCAAAGTTGAGTTCGCAAAA RT-qPCRdpsF CCGCTGCTTGGTACCATTG RT-qPCRdpsR CGCTAACAGCGATGGGAAGT RT-qPCRoxyF GGCGAAGAGGTGGTGTTACG RT-qPCRoxyR GCTCGACCAAATCATCCACATC RT-qPCRswp2079F TTAAGGCAATGGAAGCTGCAT RT-qPCRswp2079R CGTCTTTACCCGTTAATGATACGA RT-qPCRSWP_03625F GCGGAAGAAAACAGCGATAAA RNA-Seq validationSWP_03625R TGCTCACCAATACCAGCTTCTG RNA-Seq validationSWP_06535F CGAGGGCTATGCTCGTCTCTT RNA-Seq validationSWP_06535R TTGCCAGAGGCGCTTTTT RNA-Seq validationSWP_11550F TTTCACCGACTGAATACAATTTAGCT RNA-Seq validationSWP_11550R CAAGCCCAAAGACCTGCAAA RNA-Seq validationSWP_11565F ACAAGTGCTGTTGAGTTTCTAAAGGA RNA-Seq validationSWP_11565R CCACCGCACCAAATAGCATA RNA-Seq validationSWP_11570F TCAGAAACAGGGTCTACTGCAGTT RNA-Seq validationSWP_11570R ATCGACTGATTTTTGAGGCTTTAAG RNA-Seq validationSWP_18205F TGCCGAGCAAGCTTTTGC RNA-Seq validationSWP_18205R TGCTCTGCTGAAACCTTACCATT RNA-Seq validationSWP_20695F TGAGCATGCCACTGAGTATGG RNA-Seq validationSWP_20695R TTGCGCAGTACATTGAACTCAA RNA-Seq validationSWP_20720F AAACATGACCGCAGGCAGTAG RNA-Seq validationSWP_20720R GCAAGCCAGAATCGGTAACG RNA-Seq validationSWP_21995F GGCAGAGACATTCCAAACTCTAGAA RNA-Seq validationSWP_21995R GGCAGTTGATATTGCGCTTGA RNA-Seq validation

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 13

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

SUPPLEMENTAL MATERIAL

Supplemental material for this article may be found at https://doi.org/10.1128/AEM.02342-17.

SUPPLEMENTAL FILE 1, PDF file, 1.0 MB.

ACKNOWLEDGMENTSWe thank Xiaopan Ma for the valuable discussions about our data.This work was financially supported by the National Natural Science Foundation of

China (grant no. 31290232 and 41530967) and the China Ocean Mineral Resources R&DAssociation (grant no. DY125-22-04).

REFERENCES1. Oger PM, Jebbar M. 2010. The many ways of coping with pressure. Res

Microbiol 161:799 – 809. https://doi.org/10.1016/j.resmic.2010.09.017.2. Xiao X, Zhang Y. 2014. Life in extreme environments: approaches to

study life-environment co-evolutionary strategies. Sci China Earth Sci57:869 – 877. https://doi.org/10.1007/s11430-014-4858-8.

3. De Maayer P, Anderson D, Cary C, Cowan DA. 2014. Some like it cold:understanding the survival strategies of psychrophiles. EMBO Rep 15:508 –517. https://doi.org/10.1002/embr.201338170.

4. Green J, Paget MS. 2004. Bacterial redox sensors. Nat Rev Microbiol2:954 –966. https://doi.org/10.1038/nrmicro1022.

5. Zhang Y, Li X, Bartlett DH, Xiao X. 2015. Current developments in marinemicrobiology: high-pressure biotechnology and the genetic engineeringof piezophiles. Curr Opin Biotech 33:157–164. https://doi.org/10.1016/j.copbio.2015.02.013.

6. Ambily Nath IV, Loka Bharathi PA. 2011. Diversity in transcripts andtranslational pattern of stress proteins in marine extremophiles. Ex-tremophiles 15:129 –153. https://doi.org/10.1007/s00792-010-0348-x.

7. Shah J, Desai PT, Chen D, Stevens JR, Weimer BC. 2013. Preadaptation tocold stress in Salmonella enterica serovar Typhimurium increases survivalduring subsequent acid stress exposure. Appl Environ Microbiol 79:7281–7289. https://doi.org/10.1128/AEM.02621-13.

8. Garnier M, Matamoros S, Chevret D, Pilet MF, Leroi F, Tresse O. 2010.Adaptation to cold and proteomic responses of the psychrotrophicbiopreservative Lactococcus piscium strain CNCM I-4031. Appl EnvironMicrobiol 76:8011– 8018. https://doi.org/10.1128/AEM.01331-10.

9. Marietou A, Nguyen AT, Allen EE, Bartlett DH. 2014. Adaptive labo-ratory evolution of Escherichia coli K-12 MG1655 for growth at highhydrostatic pressure. Front Microbiol 5:749. https://doi.org/10.3389/fmicb.2014.00749.

10. Aertsen A, De Spiegeleer P, Vanoirbeek K, Lavilla M, Michiels CW. 2005.Induction of oxidative stress by high hydrostatic pressure in Escherichiacoli. Appl Environ Microbiol 71:2226 –2231. https://doi.org/10.1128/AEM.71.5.2226-2231.2005.

11. Kish A, Griffin PL, Rogers KL, Fogel ML, Hemley RJ, Steele A. 2012. High-pressure tolerance in Halobacterium salinarum NRC-1 and other non-piezophilic prokaryotes. Extremophiles 16:355–361. https://doi.org/10.1007/s00792-011-0418-8.

12. Bravim F, Mota MM, Fernandes AA, Fernandes PM. 2016. High hydro-static pressure leads to free radicals accumulation in yeast cells trigger-ing oxidative stress. FEMS Yeast Res 16:fow052. https://doi.org/10.1093/femsyr/fow052.

13. Xiao X, Wang P, Zeng X, Bartlett DH, Wang F. 2007. Shewanella psychro-phila sp. nov. and Shewanella piezotolerans sp. nov., isolated from westPacific deep-sea sediment. Int J Syst Evol Microbiol 57:60 – 65. https://doi.org/10.1099/ijs.0.64500-0.

14. Imlay JA. 2015. Transcription factors that defend bacteria against reac-tive oxygen species. Annu Rev Microbiol 69:93–108. https://doi.org/10.1146/annurev-micro-091014-104322.

15. Berkovitch F, Nicolet Y, Wan JT, Jarrett JT, Drennan CL. 2004. Crystal struc-ture of biotin synthase, an S-adenosylmethionine-dependent radical en-zyme. Science 303:76–79. https://doi.org/10.1126/science.1088493.

16. Conrad TM, Lewis NE, Palsson BØ. 2011. Microbial laboratory evolutionin the era of genome-scale science. Mol Syst Biol 7:509. https://doi.org/10.1038/msb.2011.42.

17. Marritt SJ, Lowe TG, Bye J, McMillan DGG, Shi L, Fredrickson J, ZacharaJ, Richardson DJ, Cheesman MR, Jeuken LJC, Butt JN. 2012. A functional

description of CymA, an electron-transfer hub supporting anaerobicrespiratory flexibility in Shewanella. Biochem J 444:465– 474. https://doi.org/10.1042/BJ20120197.

18. Chen Y, Wang F, Xu J, Mehmood MA, Xiao X. 2011. Physiological andevolutionary studies of NAP systems in Shewanella piezotolerans WP3.ISME J 5:843– 855. https://doi.org/10.1038/ismej.2010.182.

19. Jian H, Xiong L, Xu G, Xiao X, Wang F. 2016. Long 5= untranslated regionsregulate the RNA stability of the deep-sea filamentous phage SW1. SciRep 6:21908. https://doi.org/10.1038/srep21908.

20. Yang XW, Jian HH, Wang FP. 2015. pSW2, a novel low-temperature-inducible gene expression vector based on a filamentous phage of thedeep-sea bacterium Shewanella piezotolerans WP3. Appl Environ Micro-biol 81:5519 –5526. https://doi.org/10.1128/AEM.00906-15.

21. Jiang YM, Dong YY, Luo QX, Li N, Wu GF, Gao HC. 2014. Protection fromoxidative stress relies mainly on derepression of OxyR-dependent KatBand Dps in Shewanella oneidensis. J Bacteriol 196:445– 458. https://doi.org/10.1128/JB.01077-13.

22. Schutz B, Seidel J, Sturm G, Einsle O, Gescher J. 2011. Investigation of theelectron transport chain to and the catalytic activity of the dihemecytochrome c peroxidase CcpA of Shewanella oneidensis. Appl EnvironMicrobiol 77:6172– 6180. https://doi.org/10.1128/AEM.00606-11.

23. Kaya A, Mariotti M, Gladyshev VN. 2017. Cytochrome c peroxidasefacilitates the beneficial use of H2O2 in prokaryotes. Proc Natl AcadSci U S A 114:8678 – 8680. https://doi.org/10.1073/pnas.1710943114.

24. Imlay JA. 2013. The molecular mechanisms and physiological conse-quences of oxidative stress: lessons from a model bacterium. Nat RevMicrobiol 11:443– 454. https://doi.org/10.1038/nrmicro3032.

25. Calhoun LN, Kwon YM. 2011. Structure, function and regulation of theDNA-binding protein Dps and its role in acid and oxidative stressresistance in Escherichia coli: a review. J Appl Microbiol 110:375–386.https://doi.org/10.1111/j.1365-2672.2010.04890.x.

26. Xiong L, Jian H, Zhang Y, Xiao X. 2016. The two sets of DMSO respiratorysystems of Shewanella piezotolerans WP3 are involved in deep seaenvironmental adaptation. Front Microbiol 7:1418. https://doi.org/10.3389/fmicb.2016.01418.

27. van der Veen S, van Schalkwijk S, Molenaar D, de Vos WM, Abee T,Wells-Bennik MHJ. 2010. The SOS response of Listeria monocytogenesis involved in stress resistance and mutagenesis. Microbiology 156:374 –384. https://doi.org/10.1099/mic.0.035196-0.

28. Wang F, Wang J, Jian H, Zhang B, Li S, Wang F, Zeng X, Gao L, BartlettDH, Yu J, Hu S, Xiao X. 2008. Environmental adaptation: genomic analysisof the piezotolerant and psychrotolerant deep-see iron reducing bacte-rium Shewanella piezotolerans WP3. PLoS One 3:e1937. https://doi.org/10.1371/journal.pone.0001937.

29. Martinez-Garcia E, Nikel PI, Chavarria M, de Lorenzo V. 2014. The meta-bolic cost of flagellar motion in Pseudomonas putida KT2440. EnvironMicrobiol 16:291–303. https://doi.org/10.1111/1462-2920.12309.

30. Wang F, Wang FP, Li Q, Xiao X. 2007. A novel filamentous phage fromthe deep-sea bacterium Shewanella piezotolerans WP3 is induced atlow temperature. J Bacteriol 189:7151–7153. https://doi.org/10.1128/JB.00569-07.

31. Goto S, Kawamoto J, Sato SB, Iki T, Watanabe I, Kudo K, Esaki N, KuriharaT. 2015. Alkyl hydroperoxide reductase enhances the growth of Leu-conostoc mesenteroides lactic acid bacteria at low temperatures. AMBExpress 5:11. https://doi.org/10.1186/s13568-015-0098-3.

32. Smirnova GV, Muzyka NG, Oktyabrsky ON. 2000. The role of antioxidant

Xie et al. Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 14

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from

enzymes in response of Escherichia coli to osmotic upshift. FEMS MicrobiolLett 186:209–213. https://doi.org/10.1111/j.1574-6968.2000.tb09106.x.

33. Noor R. 2015. Mechanism to control the cell lysis and the cell survivalstrategy in stationary phase under heat stress. Springerplus 4:599.https://doi.org/10.1186/s40064-015-1415-7.

34. den Besten HM, Effraimidou S, Abee T. 2013. Catalase activity as abiomarker for mild-stress-induced robustness in Bacillus weihenstepha-nensis. Appl Environ Microbiol 79:57– 62. https://doi.org/10.1128/AEM.02282-12.

35. Rothe M, Alpert C, Engst W, Musiol S, Loh G, Blaut M. 2012. Impact ofnutritional factors on the proteome of intestinal Escherichia coli: induction ofOxyR-dependent proteins AhpF and Dps by a lactose-rich diet. Appl EnvironMicrobiol 78:3580–3591. https://doi.org/10.1128/AEM.00244-12.

36. Mishra Y, Chaurasia N, Rai LC. 2009. AhpC (alkyl hydroperoxide reduc-tase) from Anabaena sp PCC 7120 protects Escherichia coli from multipleabiotic stresses. Biochem Biophys Res Commun 381:606 – 611. https://doi.org/10.1016/j.bbrc.2009.02.100.

37. Choi SH, Baumler DJ, Kaspar CW. 2000. Contribution of dps to acid stresstolerance and oxidative stress tolerance in Escherichia coli O157:H7. ApplEnviron Microbiol 66:3911–3916. https://doi.org/10.1128/AEM.66.9.3911-3916.2000.

38. Chiancone E, Ceci P. 2010. The multifaceted capacity of Dps proteinsto combat bacterial stress conditions: detoxification of iron andhydrogen peroxide and DNA binding. Biochim Biophys Acta 1800:798 – 805. https://doi.org/10.1016/j.bbagen.2010.01.013.

39. Zhou A, Baidoo E, He Z, Mukhopadhyay A, Baumohl JK, Benke P,Joachimiak MP, Xie M, Song R, Arkin AP, Hazen TC, Keasling JD, Wall JD,Stahl DA, Zhou J. 2013. Characterization of NaCl tolerance in Desulfo-vibrio vulgaris Hildenborough through experimental evolution. ISME J7:1790 –1802. https://doi.org/10.1038/ismej.2013.60.

40. Harden MM, He A, Creamer K, Clark MW, Hamdallah I, Martinez KA, II,

Kresslein RL, Bush SP, Slonczewski JL. 2015. Acid-adapted strains ofEscherichia coli K-12 obtained by experimental evolution. Appl EnvironMicrobiol 81:1932–1941. https://doi.org/10.1128/AEM.03494-14.

41. Dragosits M, Mozhayskiy V, Quinones-Soto S, Park J, Tagkopoulos I. 2013.Evolutionary potential, cross-stress behavior and the genetic basis ofacquired stress resistance in Escherichia coli. Mol Syst Biol 9:643. https://doi.org/10.1038/msb.2012.76.

42. Jian H, Xiao X, Wanga F. 2013. Role of filamentous phage SW1 inregulating the lateral flagella of Shewanella piezotolerans strain WP3 atlow temperatures. Appl Environ Microbiol 79:7101–7109. https://doi.org/10.1128/AEM.01675-13.

43. Li RQ, Li YR, Kristiansen K, Wang J. 2008. SOAP: short oligonucleotidealignment program. Bioinformatics 24:713–714. https://doi.org/10.1093/bioinformatics/btn025.

44. Tarazona S, Furió-Tarí P, Turrà D, Pietro AD, Nueda MJ, Ferrer A, ConesaA. 2015. Data quality aware analysis of differential expression in RNA-seqwith NOISeq R/Bioc package. Nucleic Acids Res 43:e140. https://doi.org/10.1093/nar/gkv711.

45. Yang XW, He Y, Xu J, Xiao X, Wang FP. 2013. The regulatory role of ferricuptake regulator (Fur) during anaerobic respiration of Shewanella piezo-tolerans WP3. PLoS One 8:e75588. https://doi.org/10.1371/journal.pone.0075588.

46. Gao HC, Yang ZMK, Wu LY, Thompson DK, Zhou JZ. 2006. Globaltranscriptome analysis of the cold shock response of Shewanellaoneidensis MR-1 and mutational analysis of its classical cold shockproteins. J Bacteriol 188:4560 – 4569. https://doi.org/10.1128/JB.01908-05.

47. Edwards RA, Keller LH, Schifferli DM. 1998. Improved allelic exchangevectors and their use to analyze 987P fimbria gene expression. Gene207:149 –157. https://doi.org/10.1016/S0378-1119(97)00619-7.

Deep-Sea Adaptation and Antioxidant Defense Mechanism Applied and Environmental Microbiology

March 2018 Volume 84 Issue 5 e02342-17 aem.asm.org 15

on March 4, 2018 by S

HA

NG

HA

I JIAO

TO

NG

UN

IVE

RS

ITY

http://aem.asm

.org/D

ownloaded from